Since the 1980s I've been experimenting on and off with small scale solar power. I first bought about 1/2 a square metre of unsealed glass PV panels from an electronics surplus store in the UK. These worked quite well for several years, producing about half of their rated 12 V / 200 mA on a good day. However by the time we emigrated the output from them had dropped quite a lot so I gave them away to another experimenter in Cambridge on freecycle.
The output of the original panels had dropped in part because this is what happens with solar cells, but it was accelerated a lot because they were not sealed and they had got damp, I'd already bought another solar panel, a 12 V 28 W sealed panel from Maplin shown in the picture to the left, to supplement their output. This came with us to the Netherlands, but for three years after we moved here, it was in a box in the garage.
A few days ago I finally got it out and put it on the garage roof - flat as you see it here to try it out initially - and ran a wire into the garage. I then measured the open circuit voltage at the other end of the wire, about 18 V, and then I hooked it up to a 12 V 20 W halogen bulb in the garage. Nothing happened. Measuring, I find that there is about 0.09 V across the bulb, and the bulb itself has a resistance of 1.2 ohms, so I'm getting a mere 75 mA at 0.09 V from the cells. That's 0.007 W. That's almost nothing. Around 0.02% of the rated output.
My initial reaction was to think that the panels have a bad contact somewhere. That the bonding of wires inside is no longer good. Then I thought a bit about where the panel is located right now. As you can see, one of our neighbours has a good sized tree at the south of our garden which casts a shadow. Also, it's quite overcast weather and of course the panel is not oriented well either. While manufacturers often claim their products will work "in any light conditions" and such like, actually they need very strong sunlight to achieve full output, and it's surprising the extent to which the output is reduced by such things as what looks to a human eye as only partial shadow.
On a brighter evening a couple of days ago I held up the panel to the receding sun and measured about half the rated output, which is probably as good as can be expected at this time of year. However, unless the neighbour cuts their tree down, I don't think I'll ever generate much from a solar panel at this position, except in the summer when the sun is higher in the sky.
I also have a small solar panel on my Mango to keep the battery charged. I built a small charge regulator to go with this, as even though its potential maximum output isn't huge, it could still overcharge the battery on a good sunny day (I've also built similar regulators, but capable of dealing with more current, to work with dynamos on Mangos).
In the summer, I rarely had to use my mains electricity battery charger. Even though the orientation of the panel is never perfect due to it being mounted flat on top of the Mango, the solar panel normally provides enough current to keep the battery charged. You perhaps won't be surprised to hear that it is just about completely inoperative in winter...
Sadly, Freecycle had some problems recently so for similar ideals, you're probably better off looking here.